Prediction of drug concentration-time data in humans from animals: a comparison of three methods

The main objective of this work is to evaluate three methods to predict concentration-time data of drugs in humans in a multi-compartment system using animal pharmacokinetic parameters following intravenous administration. The prediction of concentration-time data in humans in a multi-compartment system was based on two proposed methods of Mordenti. The third method was based on the assumption that all drugs follow a single-compartment system. Ten drugs from the literature were chosen that were described by two-compartment model in both human and animals. Two-compartment model parameters (CL, V(c), V(ss), V(β), α, A, β and B) of at least 3 animals were scaled to humans and then were used to predict plasma concentrations-time data in humans. Allometrically scaled pharmacokinetic parameters from animals were also used to predict human profile using one-compartment model as a comparison. The results indicated that in a multi-compartment system, application of pharmacokinetic constants provided better prediction of concentration-time data in humans than the assumption that all drugs follow a single-compartment model. Both the proposed methods of Mordenti provided almost similar concentration-time profiles for most of the drugs. For some drugs, predicted α values were substantially higher than the observed values. This prediction error in α resulted in under-prediction of drug concentrations in distribution phase. In order to reduce the prediction error in α, Waijma's method for the prediction of α was modified which resulted in an improved prediction of concentration-time data in humans. Overall, Mordenti's proposed 2 methods and where necessary by modifying Waijma's method for the prediction of α can be used for reasonably accurate prediction of concentration-time data of drugs in humans.